Many computing devices comprise a central processing unit (CPU) and a graphics processing unit (GPU). The CPU may be configured to execute applications hosted on a computing device, such as by processing instructions of a thread for an application. The GPU may be configured to perform various tasks associated with graphics display rendering, such as rendering a user interface for display on a monitor. Because the GPU may process a substantial amount of data in a short amount of time (e.g., rendering 2 million pixels for display 60 times a second), the GPU may be configured for parallel data processing whereas the CPU may not comprise such capabilities (e.g., the CPU may be tuned for processing application code, keyboard and mouse input, internet data packets, and/or other data that may not necessitate parallel data processing). In this way, the CPU and the GPU may work together to process information, such as sensor input data (e.g., camera imagery, video input, touch input, etc.). For example, the GPU may receive sensor input data that is then provided to the CPU for processing by a code module executing on the CPU, such as an application and/or an OS code module (e.g., touch input may be processed to identify a finger gesture result, such as a scroll gesture for a user interface). The code module may send a result of the processing to the GPU for utilization in display rendering (e.g., the GPU may render the user interface based upon the finger gesture result). Because the sensor input data and/or the result is transferred between the GPU and the CPU one or more times, a user may experience latency between the touch input and viewing the rendered user interface, which may diminish the user's experience.